Does Microsoft lie about the Linux features?

Discussion in 'Linux Networking' started by totojepast, Sep 12, 2003.

  1. In comp.os.linux.advocacy, I heard Erik Funkenbusch say:
    Most monopoly crapware 'technologies' have no direct analog in the real
    market.
    That's because it has to do with interoperability, compatibility, and, yes,
    intercompatibility, things MS doesn't have.
    Depends on the thing you created, and, frankly, how well you created it.
    Any explanation that tries to explain to you how Microsoft's Kerberos
    implementation was restraint of trade is using "twisted logic", by your
    standards.

    The complaint was, at first, that they didn't document the extension. Then it
    was that they tried to hide the documentation of the extension. Then it was
    that they did hide the documentation for the extension. Then it was that the
    extension was using public fields to ensure only private compatibility. And
    that last bit turned out to be correct, because MS destroyed the market for
    Kerberos-based products by not only ensuring that the documentation wasn't
    accessible, but that it was useless. That they used the extensible abilities
    of the standard to ensure that only their implementation was interoperable
    doesn't make it OK. That they didn't succeed in establishing their
    extended-Kerberos-version as any kind of new protocol standard is hardly proof
    that they weren't trying to monopolize (by preventing 'third party'
    authentication on PC desktops from becoming viable) when they 'introduced' it.
     
    T. Max Devlin, Sep 17, 2003
    1. Advertisements

  2. In comp.os.linux.advocacy, I heard Erik Funkenbusch say:
    I can do that. Because glibc.so.6 is the most common library, at least in the
    developer community. On a consumer system, it only requires a symlink from
    that to whatever glibc-compatibile library you're using (presuming it is
    glibc.so.6 compatible) in order to allow all of your applications to work as
    if you didn't know what a symlink even was.

    I would guess, though I'm not sure because I'm not a programmer, that changing
    the makefile would be part of the recompile. Possibly all of the changing, if
    the program is "written correctly".
    Ta-daa! Which version?
    Not surprising they'd have to re-write it to be compatible with the monopoly's
    crapware.
     
    T. Max Devlin, Sep 17, 2003
    1. Advertisements

  3. Well, I assert that aliens exist, and there is proof in some book
    somewhere, just read them all and you'll see what I mean.

    Hint: References do not include pointing somewhere and saying "it's out
    there somewhere".
     
    Erik Funkenbusch, Sep 17, 2003
  4. totojepast

    Chris Guest

    I think I have to agree that NTFS is the worst, I restarted normally
    and it came back up and said I need to scan my F:\ for some reason.
    Nothing bad happend to that partition, why would I need to check my
    partition? The only reason I have ever had to run fschk is because the
    power dumped on me.
     
    Chris, Sep 17, 2003
  5. totojepast

    bgeer Guest

    Hell, they lie about their own "features"; Lying about [or is that
    lieing about...:-] features of competing software is an merely an
    extension of their SOP.
     
    bgeer, Sep 17, 2003
  6. totojepast

    Daniel Rudy Guest

    Somewhere around the time of 09/12/2003 05:48, the world stopped and
    listened as totojepast contributed this to humanity:
    FreeBSD uses MD5 hashes which is a one way encryption method for
    authenication. Let's see, you can have Kerberos4, Kerberos5, VPN, and
    IPSec in BSD. I'm sure the same is in Linux as well.

    Unix/BSD/Linux does allow the configurations of individual permissions
    to the file level. Where do you think Microsoft got the idea from?
    Microsoft products only work with other Microsoft products. The
    UNIX/BSD/Linux stuff will work with each other across different
    platforms. Solaris/AIX/HP-UX/SCO/FreeBSD/Linux. And thanks to Samba,we
    can also include Windows 95/98/98SE/Me/Xp/NT/2000/2003. Oh wait, that's
    right, you cannot run Exchage 2000 on Windows 2003 Server...So much for
    interoperability between Microsoft products.
    Tell that to SGI/SUN/IBM guys who developed various JFS implementations
    and watch them laugh you (Microsoft) right out of the forum.
     
    Daniel Rudy, Sep 18, 2003

  7. ....... Swiss security researchers announced that they'd found a way to
    crack alphanumeric passwords stored on Microsoft Windows computers in a
    mere 13 seconds, down from the previous average of 101 seconds.

    ..
    ..


    http://www.zdnet.com/anchordesk/stories/story/0,10738,2914376,00.html
     
    Vampire at Wicked Empire, Sep 18, 2003
  8. totojepast

    Wayne Throop Guest

    :: "Linux uses clear text for authentication, and does not allow the
    :: configurations of individual permissions to the file level. [...]

    : Daniel Rudy <.0123456789>
    : Unix/BSD/Linux does allow the configurations of individual permissions
    : to the file level. Where do you think Microsoft got the idea from?

    Probably via NT's VMS heritage. And while some unixen variants
    have ACLs, and while the more usual unix permissions check allows
    user and group ownership per file, the fine control of ACLs
    isn't really a "unix thing".

    And note that earlier in the thread, there was a quick backpedal,
    allowing as how linux *has* ACLs, but that anyways most backup and copy
    utilities are unaware of them, and nobody *uses* them, so that's still a
    flaw in linux (somehow).


    Wayne Throop http://sheol.org/throopw
     
    Wayne Throop, Sep 18, 2003
  9. In comp.os.linux.advocacy, Vampire at Wicked Empire
    <>
    wrote
    Ah, but watch *this* spin. :)

    (Wintrolls, if you're reading this, I am not intending this to
    be taken seriously!)

    If one forgets one password, Windows has a very convenient method
    by which one can recover it. It takes only 13 seconds...

    :)

    (Yeah, that'll make the CSO sleep better at night. Not.)
     
    The Ghost In The Machine, Sep 18, 2003
  10. totojepast

    Daniel Rudy Guest

    Somewhere around the time of 09/18/2003 10:07, the world stopped and
    listened as Wayne Throop contributed this to humanity:
    Maybe, but I wasn't really refering to ACLs. You can have file by file
    access permissions using groups and userids. It IS a pain, but it is
    possible.
     
    Daniel Rudy, Sep 18, 2003
  11. totojepast

    Bill Unruh Guest

    ]> And note that earlier in the thread, there was a quick backpedal,
    ]> allowing as how linux *has* ACLs, but that anyways most backup and copy
    ]> utilities are unaware of them, and nobody *uses* them, so that's still a
    ]> flaw in linux (somehow).

    acls would be nice, and no linux does not really have them.

    ]>
    ]>
    ]> Wayne Throop http://sheol.org/throopw

    ]Maybe, but I wasn't really refering to ACLs. You can have file by file
    ]access permissions using groups and userids. It IS a pain, but it is
    ]possible.

    Lets see, you want to give almost everyone execute permission to a file,
    you want rwx permission for yourself,
    joe is to have rx permission, but Peter is only to have have r permission.
    Now how do you do that with group and user IDS?

    There are 8 possible combinations of rwx permission, and only three
    (user, group and other) permission classes. You cannot map 8 onto 3 in a one to
    one way.

    Now how big a flaw is this? For some maybe, but that is pretty rare that you want that many
    different combinations. On the other hand four combinations. (Peter and you both with rwx,
    your supervisoris Mark and Martin with r and everyone else with no permissions) is not so rare.
     
    Bill Unruh, Sep 18, 2003
  12. totojepast

    Ed Murphy Guest

    If you're willing to fudge the location of the file, then:

    1) Create a directory that only Group1 (you, Peter, Mark, Martin)
    can access.

    2) Within that directory, create a file with 774 permissions, belonging
    to Group2 (you, Peter). Mark and Martin get world permissions.
     
    Ed Murphy, Sep 18, 2003
  13. totojepast

    Lew Pitcher Guest

    Give the file rwxr-xr-- permissions
    Give the file your own UID as the owning "user"
    Assign Joe to a unique group that does not include Peter
    Give the file Joe's group's GID as the owning "group"

    Now,
    You, as owning "user" inherit the rwx "user" permissions,
    Joe, as a member of the owning "group", inherits the r-x "group" permissions
    Peter, as neither owning "user" nor member of the owning "group, inherits
    the r-- "other" permissions.

    Having said all that, I concur that ACLs give a wider and more granular
    control over access rights than the Unix permission bits do.


    --
    Lew Pitcher

    Master Codewright and JOAT-in-training
    Registered Linux User #112576 (http://counter.li.org/)
    Slackware - Because I know what I'm doing.
     
    Lew Pitcher, Sep 20, 2003
  14. Surely you mean MP/M? CP/M was a single-user, single-tasking OS.

    Mart
     
    Mart van de Wege, Sep 20, 2003
  15. totojepast

    Les Mikesell Guest

    People keep saying this, but I've never figured out how to use windows
    ACLs to get the permissions that you usually want on a shared space
    and that the unix 'sticky bit' gives automatically. That is, how do you
    get a setting that lets anyone in the group allowed there create new files
    but subsequently only the user that created a file is allowed to delete it,
    with r/w access by others optional and controlled by the owner.
     
    Les Mikesell, Sep 20, 2003
  16. If you make a directory without delete attributes, then nobody can delete
    the file except for the files owner, the person that created it.
     
    Erik Funkenbusch, Sep 20, 2003
  17. totojepast

    James Knott Guest

    Quite so, but I'm suprised anyone remembered MP/M. However, MP/M was a
    multiuser version of CP/M from Digital Research. Now how many here
    remember the origin of DOS?
    ..
    ..
    ..
    ..
    ..
    ..
    ..
    ..
    It was originally called Q-DOS and was only intended as a development
    systems for a hardware manufacurer, Seattle Computer Products (IIRC),
    while waiting for CP/M-86. Bill Gates sold DOS to IBM, before he actually
    had any softwre to sell. He then bought Q-DOS from SCP. It was written by
    (IIRC) Tim Paterson at SCP. Since it was intended only as a development
    system, DOS was lacking many things a proper OS should have had.

    If CP/M had won out, PC software would have been much better, as it was
    designed by Gary Killdall, who was a computer science professor.

    --

    Fundamentalism is fundamentally wrong.

    To reply to this message, replace everything to the left of "@" with
    james.knott.
     
    James Knott, Sep 20, 2003
  18. totojepast

    Kenn Barry Guest

    Not bloody likely. Yes, Gary was a prof, and a damn fine
    programmer to boot, but he designed CP/M for a system with about
    _12K_ of memory. CP/M, not to mention Apple DOS 3.1 and all the
    other "operating systems" for microcomputers were minimal disk
    handlers, and PC-DOS 1.0 looked one hell of a lot like CP/M. All
    suffered from the no-multitasking legacy, and none would make a
    fit ancestor for a modern OS. You would have to go to the
    minicomputers and mainframes of that era, running things like
    UNIX and VMS, to find an OS worthy of the name.

    What was needed was to throw away those 8 bit OSes, and
    start fresh with a minicomputer OS, once peecee hardware gained
    minicomputer power.

    Kayembee
     
    Kenn Barry, Sep 20, 2003
  19. Close, but not quite an accurate description!

    The actual contracts/documents involved in the transaction
    between Seattle Computer Products (i.e., Tim Patterson) are for
    a product called 86-DOS, not Q-DOS.

    QDOS v1.0 was a "product" released by Patterson in August 1980.
    It was a fixed address version of his translation of CP/M-80 to
    8086/88 code. The 86-DOS sold to Microsoft was a different
    version of the same translation, but with relocatable
    addressing.

    Microsoft apparently did, however, use the term "QDOS" in house
    for the development version as 86-DOS was converted to what
    became MS-DOS between the time in very late 1980 when 86-DOS was
    purchased and late 1981 when MS-DOS was released.

    It appears that part of the reason for Microsoft to call it QDOS
    was an intentional, long term attempt at legally obscuring the
    fact that both they and Tim Patterson knew that running CP/M-80
    source code (which both Patterson and Microsoft had legal access
    to) through the translation software that was Patterson's bread
    and butter product, was a gross copyright violation. (Rumor had
    it that Gary Kildall was actually able to get the first version
    of PC-DOS running on an IBM-PC to print out a DRI copyright
    notice on the screen... but I've never seen that verified.)

    At any rate, Microsoft then a couple years later named even
    another product in house as "QDOS". The result was that any
    Microsoft employee could be questioned under oath and produce
    "correct" answers about "QDOS" that would not lead to either
    copyright violation disclosure or perjury charges (even though
    both would be appropriate).

    (It eventually did go to court, after Kildall sold DRI to
    Novell/Caldera. It was settled out of court in 2000.)
    To be honest, I doubt it. Kildall just simply wasn't agressive
    enough about it to make it "much better". Bill Gates wanted *very*
    badly to be the richest man in the world, Kildall just didn't
    give a ****. If he had enough money to have fun doing other things,
    he did other things. And by the time the IBM PC concept came along,
    Kildall was already well enough off to spend the rest of his life
    doing "other things". Couldn't we all be as smart and capable...
     
    Floyd Davidson, Sep 20, 2003
  20. totojepast

    Les Mikesell Guest

    Exactly right. Unix on a 386 worked better than anything else
    at the time, but we can blame its problem on another monoply.
    AT&T was charging well over $1,000 for a copy of SysV
    which put it completely out of range on the desktop, and that
    didn't include the compiler or X which were extra cost items.
    When BSDI tried to release a cheaper and more complete
    distribution, AT&T sued them. They lost, but the bad publicity
    gave windows an extra head start - and windows was still cheaper
    which is all that really mattered for sales volume. If anything like
    the current Linux versions had been around for free or less than
    windows back then there would have been no contest.
     
    Les Mikesell, Sep 21, 2003
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.